Evaluation of a Computerized Field Data Collection System for Health Surveys

Article excerpt

A customized field data collection system (FDCS) has been developed for a hand-held computer to collect and check questionnaire data. The data quality, preparation time, and user acceptability of the system were evaluated during a malaria morbidity survey in Bakau, the Gambia. Eight field-workers collected data with either the FDCS or on paper questionnaire forms in alternate weeks over a 6-week period. Significantly fewer item errors occurred with the FDCS, and by the end of the survey period interview times were significantly less with the FDCS than with the paper and pencil questionnaire.

Advanced appropriate technology may have a useful role in providing accurate and rapid information, particularly in overcoming bottlenecks in data processing, and in obviating the need for costly expertise and equipment. In developing countries this could help to improve the quality of data on health care.

Introduction

The first application of computers for the direct collection of data was for studies of telephone interviews, conducted in the 1970s 1). Recent advances in computer technology and increasing miniaturization have resulted in portable computers that now make possible computer-assisted personal interviewing (2). The facilities offered on such systems include the ability to provide range and consistency checks at the time of the interview, guided interviews with predefined routing, error messages, feedback messages, and interpretation of the information collected. Computer-assisted collection provides data in a computer-ready format, thereby eliminating the labour-intensive and error prone process of data entry, and validation. For this purpose, we designed a custom-written field data collection system (FDCS) to provide validated data entry. Following the development stage, the system underwent field evaluation within an established malaria study at the MRC laboratories in the Gambia (3). The system was compared with a traditional paper questionnaire for improvements in the quality of data, user acceptability, and robustness in daily use. Although computer-based information systems in primary health care have already been established (4), innovative applications should help overcome some of the many difficulties in planning and administering health in developing countries.

Materials and methods

The field data collection system The system hardware comprised a Psion Organiser 11 XP hand-held computer, weighing 250 g (Fig. 1), with two removable data chips. The customized software, written in OPL (a BASIC-like language), is held permanently on a 64 Kbyte EPROM chip and collected data are stored on a 32 Kbyte semi-volatile random access memory (RAM) chip. The software provides a versatile and user-friendly method of transferring questionnaires onto the hand-held computer. The construction of the questionnaire involves formulating a question prompt, an item type, and an explanatory note for each question. The item types are defined as numeric, logical, date, or character, and can be validated by range checks on numeric helds, code validity on single-coded fields, branch conditions, or consistency checks against previously recorded data. If required, records can be automatically date-, time- and identity-stamped, and access to the system restricted by password control. Customized routines can also be incorporated into the questionnaire to use child-growth reference standards and produce anthropometric data in a standardized format, or include cohort identity lists for use as a cross-reference during data entry. Several questionnaires can be defined and, when selected for data entry, supported by editing, optional help, and in-built error-message displays. Other utilities available in the data entry mode include data summary and record review. Data are transferred via a standard RS232 interface to any IBM-compatible microcomputer and this step is controlled from within the FDCS.

Study design

A 45-item questionnaire from an ongoing field study of malaria morbidity in Bakau, Gambia, was defined on the FDCS. …